Chris Evans, the Storage Architect, had an unexpected analysis of EMC's FAST announcement today on his blog. The point he makes is that with FAST only being available on Intel-based array architectures - and not the DMX product line - that EMC has put themselves at a peer level with the rest of the industry. As he says:
So what’s my point? Well, simply this; EMC have legitimised the
enterprise modular architecture characterised by V-Max. This accepts
that the future is commodity-based hardware with differentiation in
software. However, EMC are no longer the leaders in this field and are
having to play catch up.
It's an interesting angle to be sure and it will be interesting to see how EMC's new platform works out after all the bluster and FUD over FAST is long forgotten. EMC is going to busy in the years to come as they try to convert their DMX customers to v-Max customers. All vendors go through this sort of thing when they introduce new architectures and even their most loyal customers are more or less forced to consider alternatives. There is a lot of pressure on EMC all of a sudden to bring software to market that justifies the switch from the DMX -to v-MAX. FAST has failed to do that so far.
EMC's pre-announcement of FAST earlier this year was presumably done to shore up their disappointing SSD sales - a tactic that apparently did not help their enterprise business much. But having played the FAST hand so early, they had no choice but to release FAST version 0.5 before the end of the year. Will this help answer the questions about EMC's ability to develop enterprise level software for v_Max? The answer has been loud and clear - let's wait until Fast 2.0.
EMC was way too early with SSDs and they just repeated the same mistake with FAST. Instead of creating an advantage for themselves, they allowed their competitors to observe what they are doing without applying any pressure. Instead, unexpectedly, the pressure is now on EMC to do better.
Considering that 2009 was a year of backpedaling for EMC, I want to offer lessons in the art with this video:
3PAR has been designing and marketing it's agile storage for years. Recently, Hitachi started using the word "agile" in their messaging. Imitation is a form of flattery, but at least they should link to our Cloud Agile partner page when they plagiarize like that.
“Over the past two years, EMC has optimized all of its major storage
architectures to support new levels of automation required in virtual
data center environments and for the transition to private clouds.
These dynamic and agile environments require the capabilities that
EMC’s FAST technology provides for effective management and scaling.
This is an area of significant investment for EMC and we are the only
one offering this technology across high-end, mid-tier and unified
platforms.” - Pat Gelsinger, EMC President and Chief Operating Officer, Information Infrastructure Products
Online data movement is a big deal for storage and FAST puts it squarely in the mainstream of storage software. Gelsinger is right about virtual environments being dynamic and agile, but the problem with FAST is that the underlying storage does not measure up and FAST shines a spotlight on those shortcomings. If you are going to have online data movement, you need to have the ability to make the results of those data movement operations productive - without turning your storage utilization goals upside down. FAST undoubtedly will work great where EMC storage is under-utilized and very expensive, but how will it work when you are trying to reduce the cost of storage by increasing storage utilization?
It also looks like EMC is backing away slightly from their bullish position on SSDs. I was wondering how they were going to deal with their disappointing SSD sales when they finally announced FAST.
Speaking of bullish, I found this little video that I thought I'd share with you of some truly agile guys that really know their bull!
While some have wondered whether or not EMC was going to be able to get their FAST product announced this year, others have been keeping their heads down getting work done. For instance, Symantec, who today announced their Dynamic Storage Tiering feature for SSDs in their Storage Foundation products.
Of course, many probably wonder why the heck I'm writing about this seeing how 3PAR doesn't have support for SSDs in it's products yet. That's easy. A v-Max with FAST is going to be more of a science project than a production system that actually gets work done. I'm more concerned about customers buying somebody else's SSD-enabled array and using it with Symantec's Dynamic Storage Tiering than I am about FAST on v-Max.
But I'm not that concerned because the SSD market has been slow to develop - even by EMC, the company that has been leading the charge (and inventory buildup). And even though they like to point to the check boxes in the feature checklist, big wide striping (the kind that flattens disk contention problems) with a v-Max is still an exercise in storage contortionism, unlike on a 3PAR, where its the default behavior.
But all this will change over time and the cost of deploying SSDs is going to come down as enabling software for SSDs becomes available and is refined. Even EMC's software will slowly get better. Despite the performance of the SSD devices - and EMC's wishes, the race for best of breed SSD functionality is going to be much more of a marathon than a sprint.
So where's 3PAR? We've been shipping DO (Dynamic Optimization) for 4 years and today we announced the policy advisor for automating most of the tasks that used to take admins time to work through. Changing service levels for volumes just got a lot easier. Drive speed, RAID levels, additional drives - if you want to change them, you can - and it doesn't take weeks or months and piles of spreadsheets to figure it out.
3PAR customers have used Dynamic Optimization to re-balance storage volumes across multiple tiers of storage since 2005. Our Policy Advisor software utility helps them make greater use of Dynamic Optimization by analyzing volume data layouts in their 3PAR
arrays and generating the commands to move volumes to more appropriate
storage resources. Customers can choose to run those commands as a
batch process or enter them manually.
The rumor is that Copan is shuttindg down. If so, I suspect MAID storage will quickly become an afterthought now, except for a small number of customers and applications that will keep the technology on life support. The problem with MAID is that there aren't enough applications for selectively-spinning disks. Selectively spinning disk drives are more expensive than tape for archiving and are more problematic than standard disk systems for backup. That leaves applications such as video on demand, which is not a large enough market to float a serious startup these days. Thin provisioning for primary storage and dedupe for backup have become the technologies of choice for customers looking to increase the efficiency of storage.
The outcome of Copan's failure means that tape will continue to be the most prevalent technology for archiving. In turn, that means archiving software will continue to be tied to tape for some time, including all the problems inherent in maintaining metadata for it. If you are in the legal industry and hoping to see leapfrog improvements in digital discovery processes, don't hold your breath. Discovery from data on tape will continue to be a laborious task. SSDs for archiving would appear to be a slam-dunk at some distant time in the future, but the cost has to come down a long, long way first.